State-space models provide a unified approach to a wide range of problems in the time series domain. Since its creation, the software JDemetra+, which has been officially recommended by Eurostat and the ECB for seasonal and calendar adjustment of official statistics, makes a huge use of such models. The state-space framework of JDemetra+ is based on an original object-oriented design and on advanced algorithms that make it especially powerful. Even if it constitutes the kernel of many high-level routines, the state-space framework remains largely underutilized. This is basically due to the complexities of the matter and to the barrier of the programming language – Java – to access its functionalities and/or to extend them. To increase the use of the routines, we have built new modules that highly simplify in Java the creation and the estimation of univariate or multi-variate state-space models and a companion R-package for using them transparently from that well-known environment. We illustrate this new tool by some examples in R. The first one is a seasonal specific structural time series, which models series with high volatility in some periods; this is in the seasonal adjustment domain the model-based equivalent of the seasonal specific filters of X11. The second example consists in a multi-variate modelling corresponding to a complex survey design with rotating panels (Labor Force Survey in Australia). The model can be used, for instance, for the measurement of the impact of changes in the survey. Many other problems could be handled in a similar way.